SENTIMENT ANALYSIS OF DEPRESSION DISCOURSE: A CORPUS-BASED STUDY OF NOTES FROM REDDIT

Abstract

This study aims to explore the language used by the depressed community and its effects on readers' sentiments. This study examines the DNC (Depression Notes Corpus), which was scraped using Python. This corpus consisted of depression notes scraped from the social media platform Reddit between 2013 and 2023 (100 posts per year). This platform was chosen to avoid bias and maintain the anonymity of the population and sample. The purpose of this study is to conduct a detailed corpus-based sentiment analysis of notes on depression. The research question for this study is: Which lexical categories do depressed people use most, and what are the sentiments of depressed people when they write (positive, negative or neutral)? The objectives of this study are to explore the use of lexical categories and to study the impact of choices made by depressed people on others' sentiments. The depression notes were analyzed by using the corpus tool Sketch Engine. In the findings, the study revealed that the most frequent lexical category in the corpus was the verb, which suggests that depressed people want to express action. This was analyzed by examining the notes of Redditors. They mostly use adjectives to beautify their writing and to create an impact on others' sentiments. At termination, the study concludes that all positive words have positive effects on others' sentiments, negative words have negative effects, and neutral words have no effect.